The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides the agent with capabilities to read, search, and clone content from external Git repositories, which serves as an ingestion point for untrusted data.
Ingestion points: Untrusted data enters the agent context through the /repos/file, /repos/files, and /repos/grep API endpoints, as well as through git clone operations described in Procedure 2.
Boundary markers: The skill instructions do not define clear delimiters or include warnings to the agent to disregard potential instructions embedded within the retrieved code or repository files.
Capability inventory: The agent possesses significant capabilities including repository creation/deletion, commit pushing, and branch modification via the provided REST API.
Sanitization: The instructions do not mention any sanitization, filtering, or validation steps for the content retrieved from external repositories before the agent processes it.
[EXTERNAL_DOWNLOADS]: The skill references several official SDKs and packages required for integration.
Evidence: Documentation identifies the @pierre/storage package on NPM, pierre-storage on PyPI, and a Go package hosted on GitHub (github.com/pierrecomputer/sdk/packages/code-storage-go). These resources are authored by the vendor to support the skill's functionality.

code-storage